Github Awesome (@GithubAwesome) on X

Semble is code search built for AI agents, and the speed numbers are hard to ignore. It indexes a full repo in 263 milliseconds and answers queries in 1.5 milliseconds, all on CPU, no GPU, no API keys. Retrieval quality sits at 99% of a 137M-parameter transformer model. Run it as

Vol. 31 No. 2 (2026): Information Research: Artificial Intelligence (AI) in Information Science | Information Research an international electronic journal

Beyond Semantic Similarity: Rethinking Retrieval for Agentic Search via Direct Corpus Interaction

Modern retrieval systems, whether lexical or semantic, expose a corpus through a fixed similarity interface that compresses access into a single top-k retrieval step before reasoning. This abstraction is efficient, but for agentic search, it becomes a bottleneck: exact lexical constraints, sparse clue conjunctions, local context checks, and multi-step hypothesis refinement are difficult to implement by calling a conventional off-the-shelf retriever, and evidence filtered out early cannot be recovered by stronger downstream reasoning. Agentic tasks further exacerbate this limitation because they require agents to orchestrate multiple steps, including discovering intermediate entities, combining weak clues, and revising the plan after observing partial evidence. To tackle the limitation, we study direct corpus interaction (DCI), where an agent searches the raw corpus directly with general-purpose terminal tools (e.g., grep, file reads, shell commands, lightweight scripts), without any embedding model, vector index, or retrieval API. This approach requires no offline indexing and adapts naturally to evolving local corpora. Across IR benchmarks and end-to-end agentic search tasks, this simple setup substantially outperforms strong sparse, dense, and reranking baselines on several BRIGHT and BEIR datasets, and attains strong accuracy on BrowseComp-Plus and multi-hop QA without relying on any conventional semantic retriever. Our results indicate that as language agents become stronger, retrieval quality depends not only on reasoning ability but also on the resolution of the interface through which the model interacts with the corpus, with which DCI opens a broader interface-design space for agentic search.

S Banerjee (@SB434223) on X

@akshay_pachaar this is such an important point people miss with RAG embedding quality alone isn’t enough , retrieval becomes a density problem at scale as collection grow, semantic neighborhoods become crowded with “almost relevant” docs, and recall collapses which is why: - reranking

Akshay 🚀 (@akshay_pachaar) on X

A tricky LLM interview question: Your RAG system scores 90% retrieval accuracy on 5k company docs. But scaling to 500k docs drops the accuracy to just 50%, with the same embedding model and retriever. Why did this happen? The simplest answer is that more documents mean more

Saeed Anwar (@saen_dev) on X

@_avichawla The quadratic attention cost is why most production systems still chunk and retrieve instead of stuffing everything into context. Ring attention and sparse variants help but the engineering complexity they add is rarely worth it below 100k tokens.

Akshay 🚀 (@akshay_pachaar) on X

Vector DBs can't reason. Top-k similarity ranks chunks one at a time against a query. That's fine for single-hop fact lookups, and it breaks the moment a question needs information stitched across multiple chunks. That's what the FalkorDB GraphRAG-Bench results expose. The gap

Rohan Paul (@rohanpaul_ai) on X

Great survey paper on better AI memory. Modern AI needs three different memory systems: weights for slow, durable knowledge, retrieval for fresh and specific facts, and agent memory for ongoing goals, preferences, and experience. A model with only parametric memory is

Vector Database Comparison | Superlinked

Compare 47+ vector databases across features, performance, and adoption. Filter by license, languages, index types. Data sourced from VectorHub.

fly51fly (@fly51fly) on X

[IR] A Unified Model and Document Representation for On-Device Retrieval-Augmented Generation J Killingback, O Meshi, H Li, H Zamani… [University of Massachusetts Amherst & Google] (2026) https://t.co/WVmtLvwU5u